Distinguished Seminar: Chinmoy Dutta, research scientist at Lyft, Inc.
All dates for this event occur in the past.
SEMINAR
Please join us for a special seminar by Chinmoy Dutta, research scientist at Lyft, Inc.
Tuesday, September 111:30 - 2:30 p.m.
Dreese Lab, room 617 DL
“When Hashing Met Matching: Efficient Search for Potential Matches in Ride Sharing”
ABSTRACT:
We consider the problem of matching riders for sharing rides in a ride sharing platform. A daunting combinatorial challenge such platforms face is the following. Given a match pool with tens of thousands of rides, how to find potential matches very efficiently while ensuring optimality compared to exhaustive search. We design a novel algorithm for this problem using the beautiful theory of locality sensitive hashing for Maximum Inner Product Spaces (MIPS). Implementation of our algorithm could find potential matches of high quality extremely efficiently on real match pools with thousands of ride requests.
BIO:
Dutta is a research scientist at Lyft Inc. Prior to joining Lyft, he was at Max Planck Institute for Informatics, Northeastern University and Twitter Inc. He obtained his PhD in Computer Science from the Tata Institute of Fundamental Research, Mumbai and his B.Tech in Computer Science and Engineering from IIT Kanpur. Dutta's research interest lies in building products at scale for real-world problems while developing mathematical understanding of these problems. Currently, he is focused in the areas of optimization and machine learning and has been working on marketplace optimization problems like large scale online stochastic matching and autonomous vehicle prepositioning. This seminar is jointly offered by the Departments of Electrical and Computer Engineering, Mechanical and Aerospace Engineering and the Center for Automotive Research